Statistical Analysis for Education and Psychology Researchers

relationships between principals’ attitudes and involvement in staff development and

the gender, age, job related experience and academic qualifications of staff.

• Use a probability sampling strategy with a survey questionnaire. Either develop or
  choose an appropriate attitude questionnaire and follow up selected respondents with
  interviews for validation of attitude responses and additional in-depth information.

At the design stage of a survey it is useful to consider:

• purpose of the study, whether it is intended to answer specific questions or is
  exploratory;


• target population and any subpopulations of special interest;


• variables of most interest;


• appropriateness of data collection procedure;


• type of sampling and sample size;


• whether analysis is likely to be descriptive or analytical.

Example from the Literature

Cullingford (1994) surveyed 370 7–11-year-old children’s responses to advertisements
on TV. Data was collected by lengthy semi-structured interviews. Descript-ive statistics
were used to show that 75 per cent of 7-year-old children when asked whether their
favourite advertisements were true answered that they were. This was compared with 90
per cent of 11-year-old children who, in answering the same question, said that the
advertisements were not true. In the interviews children were given a chance to reflect
upon their responses and the author goes on to say that whereas the 7-year-olds’ first
response was that the adverts were true, because the products existed, their second
response (presumably on reflection) was that adverts were essentially fantasy. Similarly,
the 11-year-olds’ first response was to say the adverts were fantasy, and their second
response was to say that they were a fantasy about a real thing so were true. The author
concludes that interviews allow in-depth exploration of answers and children show a
clear shift in explaining their responses at the age of eight. The shift is in the order of
their responses, putting greater emphasis on social awareness rather than on pragmatic
facts.

1.3 Experimental Research

Experimental research is distinguished from non-experimental research by the critical
features of manipulation and control of variables to determine cause and effect
relationships. Researchers select and manipulate independent variables (sometimes
called explanatory variables) to observe the effect they have on response variables
(sometimes called dependent variables). Whereas in a general sense a variable represents
a property or characteristic of an object which varies from one object to another, when
designing a study researchers make distinctions among different types of variables
depending on the role they play in a study. For example, in one experimental study the
variable ‘degree classification’ may be the response variable or outcome variable of

Statistical analysis for education and psychology researchers 10