CONDUCTING RESEARCH IN SOCIAL PSYCHOLOGY CORRELATIONAL METHOD:

<< CONDUCTING RESEARCH IN SOCIAL PSYCHOLOGY OBSERVATIONAL METHOD

CONDUCTING RESEARCH IN SOCIAL PSYCHOLOGY EXPERIMENTAL METHOD >>

Social Psychology (PSY403)

Lesson 07

CONDUCTING RESEARCH IN SOCIAL PSYCHOLOGY

CORRELATIONAL METHOD:

Aims:

To make students understand how to conduct social psychology research using diverse scientific

methodologies.

Objectives:

Describe correlational research method

Understand causal and effect relationships in correlational research

Discuss survey method as the most commonly used method of correlational research

Understand designs used in surveys

Correlational Research

Besides describing behavior, social psychologists are also interested in learning whether two or more

variables are related, and if so, how strongly. When changes in one variable relate to changes in another

variable, we say that they correlate. Correlational research assesses the nature of the relationship between

two or more variables that are not controlled by the researcher. In studying the relationship between

children's TV viewing habits and their aggressive behavior, researchers using the correlational method do

not try to influence how much time any of the children in the study actually spend viewing violent shows.

Instead, they merely gather information on the amount of time the children spend watching such programs

and their degree of aggressive behavior, and then determine how these two variables correlate. A

correlation assesses the extent to which two variables covary

Correlation indicates?

Direction

Magnitude of a relationship

The direction and strength of the relationship between two variables are described by the statistical measure

known as the correlation coefficient (r). Strength Basis of prediction

This correlation coefficient can range from -

1.00 to +1.00 (perfect correlations). Figure 1

D ire c tio n o f R e la tio n s h ip

demonstrates positive and negative

relationships, along with zero or no

relationship.

Figure 1

Correlation & Causality

Cannot be used to determine "cause and

effect" correlation between being outgoing

and satisfied with life;

negative

no relationship

positive

Two problems in determining causality:

· reverse-causality problem (aggressive boys)

· third-variable problem: Spurious relationship

Figure 3: An example of Spurious Relationship

Correlation: Path Analysis

Social Psychology (PSY403)

A statistical technique is used to understand potential causes of correlational relationships

Figure 4 illustrates the direct relation

A n S p u r io u s R e la t io n s h ip

between parental support and getting

adjusted in college, while a mediating

· O b s e rv e d c o r re la tio n (+ d ire c t r )

variable, approach coping, is also

indirectly affecting these two links.

S h o e s iz e

M a th s k ill

Figure 4: Path Analysis

S p u rio u s c a u s a l re la tio n s h ip

A c tu a l c a u s a l re la tio n

S h o e s iz e

M a th s k ill

A ge

SURVEYS

Although studying the relationships

among variables can be done by directly

S h o e s iz e

M a th s k ill

S h o e s iz e

M a th s k ill

observing behavior or examining archived

information, it is often accomplished by

asking people carefully constructed

questions. Surveys are structured sets of questions or statements given to a group of people to measure

their attitudes, beliefs, values, or behavioral tendencies (Lavrakas, 1993; Schuman, 2002).

The

four

major

survey

techniques are face-to-face surveys,

C o rre la tio n : P a th A n a lys is

written surveys, phone surveys, and

computer surveys. The face-to-face

format provides highly detailed

· A statis tical tec hniq u e

M ediator

information and allows researchers the

Approach

use d to u n de rs ta n d

coping

best opportunity to clarify any unclear

pote ntia l ca us es of

Path C

questions. However, it is costly and

corre lation al

Path B

there is always the possibility that

relations hips

people's responses might be influenced

Change in

Parental

by the interviewer's presence. Written,

adjustment

Support

in college

phone, and computer surveys eliminate

such interviewer bias and are much less

Path A

expensive.

(direct)

An important consideration in

V alentiner et al., 1994

constructing surveys involves how

questions are asked. Survey questions usually are either open-ended or closed-ended.

An open-ended question requires a response that must have more than just a yes or no answer--

research participants provide a narrative response. A closed-ended question, in contrast, is answered with a

yes or no, or by choosing a single response from several alternatives. Closed-ended questions are the

quickest and easiest to score. In contrast, open-ended questions may provide information from respondents

that might be missed with closed-ended questions. However, open-ended responses require coding by

carefully trained judges and this is a time-consuming process.

One of the most important considerations in conducting surveys--as well as when using other

methods--is getting responses from people who represent the population as a whole. This representative

sample is often obtained through random selection, which is a procedure in which everyone in the

population has an equal chance of being selected for the sample. As long as a sample is selected randomly,

you are reasonably assured that the data will represent the overall population. However, when samples are

not randomly selected, drawing conclusions from the data can lead to serious errors.

Finally, one last problem in conducting survey research is social desirability bias, which occurs

when people respond to survey questions by trying to portray them in a favorable light rather than

responding in an accurate and truthful manner

Surveys are structured sets of questions or statements given to a group of people to measure their attitudes,

beliefs, values, or behavioral tendencies

Social Psychology (PSY403)

Types of surveys:

· Face-to-face surveys

· Written surveys

· Phone surveys

· Computer surveys.

Questions in Surveys

Types of questions:

Closed-ended questions

Open-ended questions

Examples

Closed-ended: Do you enjoy chatting online?

Response: Yes_____ No______

-How much you enjoy chatting online?

Likert Scales:

Birth control should be absolutely prohibited?

Strongly disagree

strongly agree

Birth control is a legitimate health measure?

Strongly disagree

strongly agree

Social Psychology-Lecture 2: e.g., Likert Scale

Open-ended: Why do you like (or not enjoy) chatting online?

Response ____________________________

Conclusion: Usually researchers attempt to utilize the strengths of both types; providing some alternatives,

and also giving opportunity to write response of their own choice.

Schuman et al conducted a survey in USA to identify the most important problems faced by the country;

from responses, they chose 4 problems only chosen by less than 3% of population; in another survey they

provided these four problems with an open-ended choice; 60% selected one of the rare problems;

concluded that respondents may not brain storm, are lazy, and take survey causally.

Order & format of Questionnaires

Start with easy, salient and nonthreatening questions

Avoid beginning with demographic and open-ended questionnaire

Funneling procedure: start with more general and move to more specific questions

Where appropriate, follow a chronological order, e.g., in job dates, case histories, etc.

Only use `filter' questions when unavoidable: Only use `filter' questions (e.g., if yes, go to 31, if

no, to 14) when unavoidable: it is likely that respondent may answer in such a way as to reduce the

total number of questions.

Avoid using double negatives.

In face-to-face interviews, provide cards to show the scale

Social Psychology (PSY403)

Threatening Questions

How to ask? "Did you kill your wife" (Barton, 1958)

1. The casual approach

"Do you happen to have murdered your wife"

2. The numbered card

Would you please tick the number on this card which corresponds to what happened to your wife?

The natural death

I killed her

Other (what?)

3. The everyday approach

"As you know many people have been killing their wives these days. Do you happen to have killed yours?"

4. The other people approach

Do you know any people who have murdered their wives?

How about yourself?

5. The Kinsey technique

Stare firmly into respondents and ask in a matter of fact way:

"Did you ever kill your wife?"

Remember to put such question at the end of the interview

Important considerations in conducting surveys

· Representative sample obtained through random selection

· Social desirability bias should be minimized: Social desirability bias, which occurs when people

respond to survey questions by trying to portray themselves in a favorable light rather than

responding in an accurate and truthful manner

Designs in Surveys

Survey research is carried out according to an overall plan - a research design

1. Cross-sectional design: In Cross-sectional designs, one or more samples are drawn at one point in

time; e.g., groups of different ages

2. Longitudinal design

· Trend studies/ Successive samples: Different samples of respondents from the population

complete the survey over a time period; they do not allow researcher to study changes in a

population over time. A problem occurs when the samples drawn are not comparable;

change in values of students, whether contemporary citizens are more informed than

citizens of an earlier generation.

· Cohorts studies (usually age group) typically, a cohort is an age group, but may refer to

other groups also, like people who got married during 1940.

· Panel studies: the same respondents are surveyed over time (at several points in time) in

order to examine changes in individual respondents; within subject variability is reduced;

Practice effect and drop outs are there. Moreover, if time points are more, then it is a

massive effort

Estimating sample size in surveys:

Figure 5 indicates that 2 considerations shall be in mind while deciding about number of people in one

survey:

Social Psychology (PSY403)

1. Work-out number of variables: 4 here

2. Establish minimum cell size

E s tim a tin g s a m p le s iz e r e q u ir e d

M a le

F e m a le

M a r ita l

W o r k in g

U n e m p l.

W o r k in g

U n e m p l.

S ta tu s

Age

< 20

2 1 -4 0

M a r r ie d

4 1 -6 0

60+

< 20

2 1 -4 0

N ot

m a r r ie d

4 1 -6 0

60+

T o ta l R e q u ir e d = 2 x 2 x 2 x 4 = 3 2 x 1 0 = 3 2 0

Table of Contents: