Week 1

profilemloi01
StepsinPlanningandConductingResearch.pdf

Steps in Planning and Conducting

Research

Video Title: Steps in Planning and Conducting Research

Originally Published: 2011

Publishing Company: SAGE Publications, Inc

City: Thousand Oaks, USA

ISBN: 9781483397153

DOI: https://dx.doi.org/10.4135/9781483397153

(c) SAGE Publications, Inc., 2011

This PDF has been generated from SAGE Research Methods.

SPEAKER 1: Steps in Planning and Conducting Research. Sir Isaac Newton explained gravity and

planetary orbit. Louis Pasteur said the tiny bacteria can cause disease. Benjamin Franklin claimed

that lightening

SPEAKER 1 [continued]: is electric in nature. All of these greats had something in common. They

used scientific research to learn about the world. They took the knowledge they had acquired from

others, came up with new ideas of their own, and tested them. These are all essential parts of what

is called the scientific method.

SPEAKER 1 [continued]: If we know how to conduct research, we can go about answering questions

about the nature of the environment, medicine, human beings, animals, and a host of other topics.

Conducting research helps us figure out cause and effect relationships. For example, which

environmental conditions

SPEAKER 1 [continued]: cause bees to produce the most honey? Which fertility treatments will help

the most women get pregnant? Which cancer medicine shrinks tumors with the fewest side effects?

Understanding research methods also makes us better consumers of research. If we're reading about

a study in the newspaper,

SPEAKER 1 [continued]: we'll have a better idea of whether or not we believe the results. Or if we're

advised to undergo a medical procedure, we can read the related research that has been published,

and decide whether we feel the potential benefits outweigh the risks. Let's explore the steps of

scientific inquiry that will improve your ability to draw reliable conclusions

SPEAKER 1 [continued]: in your own research, and analyze published research more critically. We'll

focus on eight steps. Choose a topic. What do you want to learn about? Generate a hypothesis. What

relationship do you suspect there may be between phenomena? Select and define variables.

SPEAKER 1 [continued]: Between which specific variables would you like to find a relationship?

Identify participants. What population are you interested in studying? Design the study. How can you

observe the phenomena in a controlled setting? Plan and conduct the research. What are the specific

steps you will

SPEAKER 1 [continued]: take to test your hypothesis? Analyze results and draw conclusions. How

can you use your data to bolster or revise your hypothesis? Share your findings. How can you tell

others what you have done, so that they can repeat and strengthen your results, or learn from your

mistakes?

SPEAKER 1 [continued]: Let's look at each of these steps individually. Choosing a Topic. The first

step in research is to choose a topic and a general research design, which means figuring out what

you want to learn about,

SPEAKER 1 [continued]: and how you can best learn about it. Some of the most common types of

research designs are observational, correlational, and experimental. Observational studies allow you

to merely examine the nature of a particular construct, that is a variable that you are interested in.

SPEAKER 1 [continued]: For example, you might be interested in determining what percentage of the

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 2 of 13 Steps in Planning and Conducting Research

population abuses alcohol. Correlational studies allow you to examine the relationship between two

or more different constructs. For example, you might want to know whether alcohol use increases as

depressed mood increases.

SPEAKER 1 [continued]: Experimental studies allow you to examine the causal effects of one variable

on another variable. For example, you might want to study whether drinking alcohol causes your

motor reflexes to become slower. Next, identify your variables.

NICOLE CAIN: A variable, sometimes known as a construct, is a special topic of interest that varies

from person to person. A person can score high on your variable, or they can score low on your

variable.

SPEAKER 1: A variable, or a construct, is a phenomenon that can be measured at higher, or lower,

levels depending upon the subjects of your study, whom or what you're observing or acting upon, and

the circumstances under which you are studying them. Some examples of variables that are studied

in the social sciences are intelligence, aggression,

SPEAKER 1 [continued]: depression, racial prejudice, and memory.

EVELYN BEHAR: There are going to be some people who are extremely intelligent, some people who

are of average intelligence, and some people who are of low intelligence. Another variable that's often

studied is aggression, or violence. Obviously in the population, you're going to have some people

who are very, very violent, some people who have maybe some tendencies towards violence, but it

inhibit it, and then some people who are not

EVELYN BEHAR [continued]: at all naturally violent.

SPEAKER 1: Next, you will need to ask a question that is scientific in nature. In other words, ask

about the relationship between one variable and another. You could choose to ask questions such

as, what is the impact of depression on family relationships. What is the impact of racial prejudice on

juror perceptions?

SPEAKER 1 [continued]: Or what is the impact of anxiety on memory? When creating your study, it

is important to choose a topic that you can actually measure. If you were a botanist, it would be fairly

easy to measure the effect of watering a plant on the plants growth by controlling the amount of water

you give the plant,

SPEAKER 1 [continued]: and physically measuring how big the plant gets. Other studies, such as

those in the social sciences, can be more complicated.

NICOLE CAIN: Some populations, or constructs, need special considerations in order to be

measured. One example of that would be if you were interested in examining how brain activity plays

a role in depression. You would first need to make sure that you have a way of measuring brain

activity before starting the study.

EVELYN BEHAR: If you're interested in studying genetic transmission in schizophrenia, you obviously

are going to need a way to measure genetic transmission. So you're going to have to have access to

some sort of DNA testing technology, in order to ultimately answer your question.

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 3 of 13 Steps in Planning and Conducting Research

NICOLE CAIN: Another example of that would be if you were interested in racial attitudes of jury

members. You would want to have participants who are actual jurors, but that might not always be

possible. So you would need to create what's called an analog situation.

EVELYN BEHAR: An analog situation is essentially when you ask your participants to pretend that

they are in a particular situation. So you might ask your participants to imagine that they are members

of a jury, and to listen to the case before them, and then to answer a series of questions. What you

might want to do, in these cases,

EVELYN BEHAR [continued]: in order to establish what's a much more believable analog situation,

is do something like set up your laboratory to look like a courtroom, so you could have a judge, you

could have a bailiff, you could have attorneys. And you could really, if you put enough money and

time into it, you could really make your laboratory look very realistic.

SPEAKER 1: To summarize, you'll need to make sure that you have the technology needed to

measure your variables, as well as a setting that is conducive to accurate replication of participants'

behavior. Before you settle on a research question, you'll want to take the time to read the scientific

literature to make sure that your question hasn't been answered

SPEAKER 1 [continued]: in past studies, and that there is a good theoretical basis for asking the

question. Generate a Hypothesis. By the time you finish reading the scientific literature about your

topic, you'll probably

SPEAKER 1 [continued]: have an idea as to what impact you think your first variable has on your

second variable. This is a hypothesis.

NICOLE CAIN: A hypothesis is a prediction about how your variables of interests will relate to

each other. Hypotheses should be based on previous research. It's not good enough to just take

a wild guess about how your variables would relate to each other, you need to look at what other

researchers have found to be true.

EVELYN BEHAR: So let's say, for example, that you know based on past research that when people

are anxious, they tend to have poor memory skills. And you want to now come along and run an

actual experiment to look at the causal relationship between these two things. So you want to ask the

question, if people are anxious will that

EVELYN BEHAR [continued]: cause them to have poor memory.

SPEAKER 1: A hypothesis takes the form of an if then statement. In a correlational study, in which

we are just observing, you may predict that if a certain condition exists, then it is more likely, or less

likely, for some other condition to exist. In an experiment, you will be looking for cause and effect.

SPEAKER 1 [continued]: So your hypothesis will be along the lines of, if a certain action or

circumstance is imposed, then a certain outcome will take place. Select and Define Variables.

SPEAKER 1 [continued]: Selecting and defining your variables is one of the most important steps in

the research process, because choosing good variables, and good definitions of those variables, may

make the difference between finding interesting results, and not finding anything useful.

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 4 of 13 Steps in Planning and Conducting Research

NICOLE CAIN: All experiments are made up of two different types of variables, independent

variables, and dependant variables. The independent variable is a variable that is active in your

research study. It's the variable that you, as the experimenter, manipulate during the course of your

study. So for example, you could be interested in studying

NICOLE CAIN [continued]: the effects of mood on memory. So you can bring people into the lab, and

induce a mood in them-- a positive mood or a negative mood-- maybe through having them watch

sad or happy movie clips. Your independent variable would be the mood state that you were inducing

in your participants. In contrast, the dependent variable is a passive variable.

NICOLE CAIN [continued]: It's the variable that the independent variable acts upon. It's the variable

that you're measuring as part of the study. So to use our example, the recall of the list of words, or

the number of words that they can remember, is your dependent variable. All experiments must have

at least one independent variable that would have at least two different levels.

NICOLE CAIN [continued]: In our example, it would be the positive versus the negative mood. In

addition, all studies need to have at least one dependent variable.

EVELYN BEHAR: So just to recap, the independent variable is the active variable, it's what you

manipulate as an experimenter. And the dependent variable is the passive variable, it's the thing that

gets measured. It's the thing that is acted upon.

NICOLE CAIN: All variables can be expressed in two different ways, conceptually or operationally.

When you define a variable as conceptual, it's the general more abstract way of thinking of your

variable. When you want to be more specific, you look at the operational definition of your variable.

This is the more specific and concrete way of thinking about how you're going to measure or

manipulate

NICOLE CAIN [continued]: your variable.

EVELYN BEHAR: Intelligence, which is a variable, is really a conceptual variable. It's kind of abstract.

You want to be able to measure it in some concrete way, and you're going to operationalize it by

perhaps giving people an intelligence test.

SPEAKER 1: Remember, conceptual variables are general. Operational definitions are specific. Your

independent variable, which you manipulate, can be applied at two or more levels. If you include only

two levels-- for example, if you have participants in your study on sleep deprivation--

SPEAKER 1 [continued]: get no sleep or a full night's sleep, you can find only a linear relationship

between the levels. In this case, it will appear that there was a clear effect of sleep deprivation on

depressed mood. Participants who got no sleep at all are in a depressed mood. And participants who

got a full night's sleep are in a fine mood.

SPEAKER 1 [continued]: If you include three or more levels-- for example, no sleep, four hours of

sleep, and eight hours of sleep-- then you may end up with a non-linear relationship, which could tell

you something more complex about the relationship between sleep and depression. In this case, it

appears that getting no sleep and getting a full night's sleep may both lead to a fine mood,

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 5 of 13 Steps in Planning and Conducting Research

SPEAKER 1 [continued]: while participants who got only four hours of sleep are in a depressed mood.

Identify your Participants. The next step is to choose your participants. Who will be part of your study?

SPEAKER 1 [continued]: You may be interested in learning about everyone in the world, but more

likely you'll choose a more specific population.

NICOLE CAIN: There are many different populations that researchers can draw from. Some of the

examples would include high school students, college students, patients in mental health setting,

or prisoners. Once you've identified your population, you must select your sample. Your sample is a

subset of the population that you want to study.

EVELYN BEHAR: As researchers, we're interested in potentially many different populations. So for

example, one researcher may be interested in prisoners, another researcher maybe interested in

psychiatry in patients, yet another researcher might be interested in infants. Let's say that I am

interested in prisoners. This is the population that I'm interested in.

EVELYN BEHAR [continued]: Now once I have established that, I need to select my sample, which

is a subset of the population. Ideally it would be lovely if I could go out there and measure every

single prisoner in the world, but obviously that's not realistic. It's going to be too expensive, it's going

to take up too many resources. So instead, I'm going to select a sample. Say I select a sample of 200

prisoners.

EVELYN BEHAR [continued]: One thing that I need to take into consideration is the idea of selecting

a random sample. And what that means is that every single person in that population of prisoners--

that means every prisoner in the world-- has an equal chance of ending up in my study, ending up in

my sample. That's a random sampling. And this is an ideal in research.

EVELYN BEHAR [continued]: This is something that we strive for, but we often can't actually get

there, and this is why. Let's say that I live in Pennsylvania, and around me there are 10 different

prisons in the state that I could go and measure prisoners. And that's great, and that's probably what

I'm going to end up doing as a researcher in Pennsylvania, but when I get my results

EVELYN BEHAR [continued]: we have a potential problem. The potential problem is that I may be

answering questions about prisoners in Pennsylvania. I may not be answering a question about

prisoners all over the world. Perhaps there is something different about prisoners in New York, or

Florida, or California relative to prisoners in Pennsylvania.

EVELYN BEHAR [continued]: So even though I'm going to strive for getting a random sample from

my study, it's probably unlikely that I'm actually going to be able to get a truly random sample in my

investigation. The truth is, all research investigations are limited in terms of which the sample we're

selecting. And if you think about it in the most simple term,

EVELYN BEHAR [continued]: even just picking up the telephone, and calling a potential participant

already ensures that you don't have a truly random sample. That's because there are some people in

the world who don't have a telephone. So by definition, you are systematically excluding those people

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 6 of 13 Steps in Planning and Conducting Research

who don't own a telephone, perhaps people in very, very rural areas.

SPEAKER 1: Select an Appropriate Design. At this point, you have a hypothesis and a population.

You know what you're studying, and you know who will participate in your study. How will you conduct

your study? You are now ready to select design features that

SPEAKER 1 [continued]: will help you find answers. There are two main decisions you need to make.

The first decision is whether to have more than one independent variable. If you choose to have only

one independent variable, this is called a one-way design. This type of experiment is relatively simple

and straightforward.

SPEAKER 1 [continued]: Remember that even in a one-way design, you can include more than two

levels of the independent variable, so that you can draw nonlinear conclusions. A factorial design has

more than one independent variable. It is usually beneficial to use a factorial design, because it is

very rare for only one construct, or variable,

SPEAKER 1 [continued]: to be influencing a dependent variable.

EVELYN BEHAR: It's really important to try, if you can, to have more than one independent variable

in your study, and here is why. Let's say that you are interested in the effects of sleep deprivation on

mood the next day. So we all know that when we've been sleep deprived, maybe we can be a little bit

crabby the next day, or a little bit overly sensitive. However it's unlikely that sleep deprivation

EVELYN BEHAR [continued]: is the only thing that's impacting mood the next day. It's probably

the case that there are lots of variables that could impact your mood the following day. So ideally,

in addition to sleep deprivation, you might want to also have a measure of people's relationship

problems, maybe their food intake, because we know that these are also variables that

EVELYN BEHAR [continued]: can impact the next day's mood. So again, just to recap, you want to

make sure if you can, whenever possible, to not only have one independent variable in your study,

but to have multiple ones, because in the real world we're not just affected by only one variable. We're

affected by lots of variables in our lives.

SPEAKER 1: If you examine two or more variables, you'll get a more complete picture of what is

impacting your dependent variable. If you're using a factorial design, you'll need to keep track of which

level of each independent variable is being applied in each case. The method for keeping track of

these combinations is called factorial notation.

SPEAKER 1 [continued]: For example, you may choose to have three independent variables-- sleep

deprivation, caffeine intake, and life stress. You may have three levels of sleep deprivation, two levels

of caffeine intake, and three levels of life stress. This would be called a 3 by 2 by 3 factorial design.

SPEAKER 1 [continued]: There are three numbers, because there are three independent variables

in the study. Each of these numbers tells you how many levels exist, within a given independent

variable. The first three tells you that there are three levels of sleep deprivation-- no sleep, four hours

of sleep, and a full night's sleep.

SPEAKER 1 [continued]: The two tells you that there are two levels of caffeine intake-- one cup or

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 7 of 13 Steps in Planning and Conducting Research

three cups. The second three tells you that there are three levels of life stress-- low, medium, and

high. In this example of factorial design, you'll have 18 unique conditions, or cells.

SPEAKER 1 [continued]: You find this number by multiplying the number of levels within each

independent variable. Here you multiply 3 by 2, which is 6, and then multiply that by 3, which brings

you to 18. If there were more independent variables, you would continue to multiply by the next

number of levels. The product-- which is 18 in our example--

SPEAKER 1 [continued]: tells you the number of cells in the experiment. Using a factorial design

is more complicated, but it allows you to ask more realistic questions, and create a scenario that is

closer to the real world, where more than one variable affects the dependent variable.

SPEAKER 1 [continued]: The other major decision you need to make in designing your experiment is

whether participants will serve in one, or more than one, cell of the study. In between-subject designs,

each participant serves in only one cell of the experiment. For example, in the sleep deprivation study,

you would need 360 different participants

SPEAKER 1 [continued]: in order to have 20 participants in each cell. That is 20 participants times

18 cells. On the other hand, if your plan is to have participants serve in more than one cell, you have

within-subjects design. In our example, you might take the life stress variable, and make it within-

subjects variable.

SPEAKER 1 [continued]: You would then expose each participant to each of the three stress levels,

and measure their mood after each one. Plan and Conduct Research. Once you have determined

who your participants are,

SPEAKER 1 [continued]: and what kind of study you are conducting, you can begin the hands on

creation of the experiment. This means setting up your laboratory, so that it is appropriate for your

study, which sometimes means transforming it into something that no longer seems like a laboratory

at all. An important concept in creating your study

SPEAKER 1 [continued]: is experimental realism.

NICOLE CAIN: Experimental realism means that you want to try to set up your laboratory as close as

possible to a real world. So for example, if you're interested in looking at attitudes of jury participants,

you would want to actually take the time and effort to set your laboratory up, so that it looks like an

actual court room. It gives the participants in your research study

NICOLE CAIN [continued]: a chance to act naturally, and act as though they were actually in a court

room, giving you more real life data.

EVELYN BEHAR: You might not get there 100%, but you can at least increase the likelihood that

you're going to get individuals, participants in your study, actually behaving as they normally would.

SPEAKER 1: Another thing to remember when you're running an experiment is that you must

randomly assign your participants to the various conditions. Without random assignment, you do not

have a true experiment. Non-random assignment could lead to biased assignment.

EVELYN BEHAR: One of the hallmarks, if not the hallmark of an experimental study, is the idea of

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 8 of 13 Steps in Planning and Conducting Research

random assignment. So when you are creating or designing an experiment, you want to make sure

that you are randomly assigning your participants to the different conditions of your experiment, and

you want to do it in a way that is not at all biased.

EVELYN BEHAR [continued]: So very old-fashioned, but very effective way, is to literally flip a coin,

and decide is the participant going to end up in condition a or condition b. And you want to be very

careful. You don't want to let your emotions get in the way. So let's say that the first participant who

arrives for your study is a woman named Mary,

EVELYN BEHAR [continued]: who is clinically depressed and here for a study comparing cognitive

behavioral therapy for depression to a wait list comparison condition. And you flip your coin and it

lands on tails, and that means that Mary is about to go into the weightless condition. And you haven't

even told Mary, yet but Mary is all ready.

EVELYN BEHAR [continued]: She's crying, she's weepy, she's telling you about all of her life

problems that go along with her depression. And it starts to pull at your heartstrings a little bit. And

you say to yourself, I just don't have the heart to put Mary in the weightless condition. I'm going to

save the weightless condition for someone maybe who's suffering a little bit less. And I'm going to go

ahead and put Mary in the active treatment condition, because I really care about Mary, and I like her,

EVELYN BEHAR [continued]: and I want her to get better. Even though you're being very sensitive,

you have broken one of the cardinal rules of experimentation, which is to stick to the random

assignment plan. So when somebody comes in and you flip that coin, you absolutely, without any

exceptions, you must put them into the condition

EVELYN BEHAR [continued]: to which they've been assigned.

SPEAKER 1: Now you're ready to run the experiment and collect the data. This is the crux of the

matter, though it is crucial that you complete the previous steps so that your study yields credible

information, and the upcoming steps so that you can share what you've learned with others.

NICOLE CAIN: So it's important for you to standardize your entire studies, so that all participants in

your study undergo the exact same condition. So for example, if you were interested in studying

personality traits of criminals versus non-criminals, you want to make sure that you're holding all

things constant in your study, and that you're treating both groups equally.

NICOLE CAIN [continued]: This will ensure that there are no differences between the two groups,

except for your experimental manipulation.

EVELYN BEHAR: Every person who walks into your laboratory for that study, no matter which

condition they're in, gets the thing treatment and lots of different levels, except for that one

independent variable. So how do you make sure that you're treating everybody exactly the same?

You might want to have a script that the experimenter follows. You want to make sure that everybody

EVELYN BEHAR [continued]: is going into the same room. Also it would be helpful if the research

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 9 of 13 Steps in Planning and Conducting Research

assistant, or the experimenter, were naive about the whole purpose of the study. So when you have

this person working for you, and meeting with all of the different participants who come through the

door, that person should not know what the hypotheses of the study are.

SPEAKER 1: Analyze Results and Draw Conclusions. Data analysis can be very complex, and

becoming an expert requires many years of instruction. But there are some basics you should know.

Before you begin an in-depth analysis,

SPEAKER 1 [continued]: you may want to create a rudimentary graph, so that you can see if any

patterns jump out at you. Your independent variable will be along the x-axis, and your dependent

variable will be along the y-axis. One of the first things you want to look for is central tendency, or

participants' typical performance on your variables of interest.

SPEAKER 1 [continued]: There are three measures of central tendency. The first measure of central

tendency is the mean, this is the average of a distribution, usually calculated separately for unique

conditions of an experiment, so that can later be compared. Each mean is the average of all the

scores for a given condition, or set of conditions.

SPEAKER 1 [continued]: They are added up and divided by the number of participants in that group.

The second measure of central tendency is the median. Unlike the mean, which takes the average

of all of the scores, the median is the middle number in a distribution of scores. If there are 25

participants in a group,

SPEAKER 1 [continued]: and you write out their scores in ascending order, the median will be

whatever score appears 13th, or right in the middle. The third measure of central tendency is the

mode. The mode, is the most commonly recorded value in a distribution of scores.

SPEAKER 1 [continued]: For non-experimental research, though you are not manipulating the

variables, you can investigate more than one variable, and analyze your data for correlations. These

correlations do not tell you about cause and effect, but they do tell you about the relationship between

two variables. Correlations range from negative 1 to positive 1.

SPEAKER 1 [continued]: A positive number means that there is a positive relationship between the

two variables. In other words, as one variable increases, so does the other. A negative number

means that there is a negative relationship between the two variables. As one increases, the other

decreases.

SPEAKER 1 [continued]: The size of the number tells you the magnitude of the correlation. The closer

the value is to the extremes, that is to positive 1 or negative 1, the stronger the relationship is between

the two variables. For example, 0.9 stronger than 0.3. 3 Negative 0.4 is stronger than 0.2,

SPEAKER 1 [continued]: even though one is negative and the other is positive. When we conduct

experiments, we're trying to find cause and effect by manipulating the independent variables.

Analyzing the data from experimental studies differs depending on the experimental design. Let's first

look at one-way designs,

SPEAKER 1 [continued]: experiments with just one independent variable. To understand the results

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 10 of 13 Steps in Planning and Conducting Research

of your study, you'll statistically compare the means of the different groups. In our earlier example--

examining the effect of sleep deprivation on mood-- we had three levels of sleep, no sleep, four hours

of sleep,

SPEAKER 1 [continued]: and a full night's sleep. Let's assume you measured depressed mood using

the Beck Depression Inventory , or BDI. You might find the following. Participants who got no sleep

had a mean BDI of 26. Participants who got four hours of sleep had a mean BDI of 18.

SPEAKER 1 [continued]: Participants who got eight hours of sleep had a mean BDI of eight. Here

you would run a T-test to statistically test for differences between the three levels of sleep, and the

results of the test would tell you whether those three levels yielded significantly different values on

the dependent variable-- that is,

SPEAKER 1 [continued]: depressed mood or BDI. This is necessary in order to draw conclusions,

which we will talk about shortly. For factorial designs, experiments with more than one independent

variable, you will run an analysis of variance, or ANOVA. The ANOVA allows us to answer questions

SPEAKER 1 [continued]: about the effects of each of the independent variables, and the possible

interaction between or among them. Let's consider a 2 by 2 factorial design. This is the most common

type of design in research studies, and it will enable you to walk through the ANOVA process.

Running an analysis of variance will allow

SPEAKER 1 [continued]: you to answer three questions. One, is there a main effect of the first

independent variable? This ignores the influence of the second independent variable. Two, is there

a main effect of the second independent variable? This ignores the influence of the first independent

variable. And three, is there an interaction between the two

SPEAKER 1 [continued]: independent variables? This takes both independent variables into

consideration. In our sleep deprivation example, using just two levels of sleep-- 0 hours and 8 hours--

and two levels of stress-- low and high-- you would ask these three questions. One, is there a main

effect of sleep deprivation?

SPEAKER 1 [continued]: Perhaps you'll find that participants who got no sleep at all show higher

depressed mood than participants who got 8 hours of sleep. Two, is there a main effect of stress

level? Perhaps you'll find that participants who underwent high levels of stress show higher

depressed mood, than participants who

SPEAKER 1 [continued]: underwent low levels of stress. Three, is there an interaction between sleep

deprivation and stress level? You could find that participants who receive no sleep showed more

depressed mood if they underwent high levels of stress, than if they underwent low levels of stress.

SPEAKER 1 [continued]: But that those who received eight hours of sleep showed more depressed

mood if they underwent low levels of stress, than if they underwent high levels of stress. Now you're

ready to draw your conclusions.

EVELYN BEHAR: At the end when you've actually analyzed your data and come up with your results,

what you want to do is go back to that initial hypothesis, or prediction, and compare them. You want to

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 11 of 13 Steps in Planning and Conducting Research

see well, I made this prediction, I posed this hypothesis, these were my results. Do my results support

the hypothesis or do they refute the hypothesis?

NICOLE CAIN: If your original hypothesis has been refuted, you want to think about why that might

be, and also think about how that impacts the theory that your hypothesis was drawn around. If your

hypothesis is supported, you want to think about replicating your results. You may have actually found

this result by chance. This happens, this is statistically possible.

NICOLE CAIN [continued]: So you want to make sure that you can find the same result a second time.

If you can actually replicate your study using the same methods, and a different set of participants,

with even a different set of experimenters, this gives you a lot more confidence that your result is

accurate.

SPEAKER 1: Share Findings. Now you're ready to share your new knowledge with the world. Here

are some guidelines that will help you figure out what you need to include in your research report.

First, you need to write an introduction describing the theoretical background of your study, past

evidence that

SPEAKER 1 [continued]: supports your hypotheses, and how this idea developed logically, based

on past studies and existing theories. Next, you'll explain how you conducted your experiment.

How many participants were included, and how did you select them? What are their demographic

characteristics-- their age,

SPEAKER 1 [continued]: race, or ethnicity, et cetera. You'll need to share your experimental design.

For example, you can explain that you created a 2 by 2 between-subjects factorial design, and

discuss the independent and dependent variables. Then you'll explain your procedure by giving a

step-by-step explanation of what

SPEAKER 1 [continued]: participants did in the experiment, any special equipment used to collect

data, how variables were operationalized-- that is defined in a way that is measurable-- and similar

details. Of course, you also want to share your results. Include all of your data analyses. Finally

discuss your results in light of existing research.

SPEAKER 1 [continued]: How this adds knowledge to the world. What the limitations of the study

were, and how this impacts real-world practices. You may also make suggestions for future research.

Conclusion.

SPEAKER 1 [continued]: Now that you've learned about planning and conducting research, you can

begin to think about what you would like to add to the world of scientific investigation. Remember,

here are the basic steps. Choose a topic, read existing research before you decide what you will

study, generate a hypothesis-- your if then prediction-- select

SPEAKER 1 [continued]: and define independent and dependent variables, and operationalize them,

identify participants. Remember, you're finding a sample within a population. Design the study,

include whether you'll have one independent variable, or more than one. Plan and conduct the

research.

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 12 of 13 Steps in Planning and Conducting Research

SPEAKER 1 [continued]: Analyze results and draw conclusions. And finally, share your findings so

that others may learn from your research. Good luck and see you in the laboratory.

SAGE

2011 SAGE Publications, Ltd. All Rights Reserved.

SAGE Research Methods Video

Page 13 of 13 Steps in Planning and Conducting Research

  • Steps in Planning and Conducting Research