Social Work Research Questions and Answers

Answer: A

Mixed-method research combines Quantitative and qualitative methods.

Mixed-methods designs, developed by writers such as John Creswell, collect and analyse both numerical data (surveys, measurements, statistics) and qualitative data (interviews, observation, texts) within one study, so that each type compensates for the limits of the other. Typical designs are convergent, explanatory sequential and exploratory sequential.

The other pairs are not what the term means. A survey and a case study are particular designs, and a survey is usually quantitative. Participatory research is an approach that can use either kind of data. 'Qualitative and survey' mixes a type of data with a single method.

Answer: D

In a normal distribution, Mean, Median and Mode are equal.

The normal curve is perfectly symmetrical and has a single peak. The highest point (the mode) lies at the centre, half the values fall on each side of that centre (so it is also the median), and the balancing point of the values (the mean) is at the same place.

Options A, B and C describe skewed distributions. In a positively skewed distribution the long tail is on the right and mean is greater than median, which is greater than mode; in a negatively skewed one the order reverses. None of these applies to a symmetric normal curve.

Answer: A

A hypothesis is not necessary in an Exploratory research design.

Exploratory research is done when little is known about a problem. Its purpose is to gain familiarity, clarify concepts and develop insights, and so it often ends by producing hypotheses for later studies. It does not start from one.

Explanatory (causal) and correlational designs exist to test whether variables are related or one causes another, so they are built around hypotheses. Descriptive studies set out to describe a situation accurately and usually work from clearly defined questions and expectations. Of the four, only exploratory research is defined by not needing a hypothesis in advance.

Answer: D

The social distance scale was developed by Bogardus.

Emory S. Bogardus published the scale in 1925. It measures how close a relationship a person is willing to accept with members of another group, through a series of items from 'would admit to close kinship by marriage' down to 'would exclude from my country'. It is a cumulative scale: accepting a close relationship usually implies accepting the more distant ones.

Rensis Likert developed the summated rating scale in 1932. Emile Durkheim was a founder of sociology and did not devise attitude scales. 'Beacon' is not the author of any well-known scale.

Answer: B

The researcher used Multistage sampling.

In multistage sampling the sample is drawn in successive stages, each from within the units chosen at the stage before. Here districts were chosen first, then talukas within those districts, then villages within those talukas, and finally women respondents from the Gramsabhas of those villages. That chain of levels is the defining feature.

Stratified sampling divides the population into strata, such as caste or income groups, and samples from every stratum, rather than narrowing down through levels. Quota sampling fills fixed numbers non-randomly. Sequential sampling keeps adding cases until a decision can be reached, with no fixed sample size.

Answer: C

A t-test tests the statistical significance of a Difference.

The t-test asks whether the difference between two means is larger than chance would produce: a sample mean against a known value, two independent groups, or the same group before and after. The t statistic divides the difference by its standard error:

\[ t = \frac{\bar{x}_1 - \bar{x}_2}{SE_{\bar{x}_1 - \bar{x}_2}} \]

Association between categorical variables is tested with chi-square. Correlation is measured with a coefficient such as Pearson's r (whose significance can be checked separately). Comparing variances, or several group means through them, uses the F-test and ANOVA.

Answer: C

Intervention research mainly helps social workers assess the effectiveness of social work practice.

Intervention research designs, tests and evaluates the methods, programmes and techniques that social workers use with individuals, groups and communities. Its question is whether a given intervention produces the intended change, which is the basis of evidence-based practice.

Theories, principles and philosophy guide practice, but they are not what an intervention study measures. It is the practice, the actual helping activity, whose effects are tested.

Answer: B

The test that is not parametric is the Wilcoxon test.

Parametric tests assume that the data come from a population with a known distribution, usually normal, and use interval or ratio data. Non-parametric tests make no such assumption and work on ranks or signs. The Wilcoxon signed-rank and rank-sum tests are non-parametric, the rank-based alternatives to the t-test.

Pearson's coefficient of correlation, the F-test (as in analysis of variance) and Student's t-test all assume normally distributed interval data, so they are parametric.

Answer: C

The embedded design is a mixed-method research design.

In the classification of Creswell and Plano Clark, the embedded design is one of the mixed methods designs: one kind of data, quantitative or qualitative, plays a supporting role within a study based mainly on the other kind, for example qualitative interviews embedded within an experiment.

Because it combines both kinds of data by definition, it cannot be a purely quantitative or purely qualitative design. A case study is a separate strategy of inquiry, not the category the embedded design belongs to.

Answer: D

It is the null hypothesis.

The null hypothesis (H0) states that there is no relationship or no difference between the variables, which denies the relationship the research hypothesis proposes. The statistical test is applied to the null hypothesis, which is either rejected or not rejected.

A directional hypothesis predicts the direction of a relationship, and a complex hypothesis involves more than two variables, so both assert a relationship. 'Statistical hypothesis' is a general name for any hypothesis stated in testable statistical form, not specifically the one that denies the relationship.