Designing your research project

Release the Kraken!

Esteban Montenegro-Montenegro, PhD

Psychology and Child Development

Aims in this lecture

  • Brief refresher on research designs.

  • What is a dependent variable?

  • What is a independent variable?

  • Hypothesis creation.

The big picture again

PSYC4310researchProject

Big research families

Alternative research designs (Creswell & Creswell, 2017)
Quantitative Qualitative Mixed Methods
Experimental designs Narrative Research Convergent
Non-experimental Phenomenology Explanatory sequential
Longitudinal Designs Grounded Theory Exploratory sequential
Ethnographies Complex designs with embedded core designs
Case Study

Big research families II

  • Survey research: provides a quantitative or numeric description of trends, attitudes, or opinions of a population by studying a sample of that population. It includes cross-sectional and longitudinal studies using questionnaires or structured interviews for data collection—with the intent of generalizing from a sample to a population.

  • Experimental research: seeks to determine if a specific treatment influences an outcome. The researcher assesses this by providing a specific treatment to one group and withholding it from another and then determining how both groups scored on an outcome.

    • True experiments
    • Quasi-experiments
      • Single-subject designs

Big research families III

Qualitative designs

  • Narrative research: The information retold or restoried by the researcher into a narrative chronology. Often, in the end, the narrative combines views from the participant’s life with those of the researcher’s life in a collaborative narrative

  • Phenomenological research: the researcher describes the lived experiences of individuals about a phenomenon as described by participants.

  • Grounded theory: is a design of inquiry from sociology in which the researcher derives a general, abstract theory of a process, action, or interaction grounded in the views of participants.

  • Ethnography: is a design of inquiry coming from anthropology and sociology in which the researcher studies the shared patterns of behaviors, language, and actions of an intact cultural group in a natural setting over a prolonged period of time.

  • Case studies: in-depth analysis of a case, often a program, event, activity, process, or one or more individuals.

Let’s focus on causality for a while

What is first A or B ?

  • Causality means that we would expect variable X to cause variable Y.
    • For example: Does low self esteem cause depression? How do we know?

Let’s take a look at some spurious correlations:

Experiments help us!

  • Can we know if A causes B with a survey?

  • Can we know if A causes B conducting an experiment?

  • We can manipulate a variable and observe what happens afterwards, but it is good enough?

  • Do we need something more on our design?

Experiments help us! II

  • Pure experiments need a control group to account for counterfactual information, this also helps to rule out possible confounding variables.
    • Example: how would you measure the effect of physical activity on cardiovascular fitness? What would be a good experimental design?

Can we assume causality in survey designs ?

  • In survey designs we cannot manipulate the independent variable, but some researchers claim that is possible to make causal inferences when you conduct a longitudinal study.
  • In longitudinal studies you satisfy the temporal requirement, you could evaluate if X = independent variable happens before Y = dependent variable.
    • For instance: You could measure a baby’s weight every month and evaluate how many times the baby is breastfed. But, do we need counterfactual information?

But we haven’t defined cause, and effect III

What is an effect?

  • An effect is better define if we have a counterfactual model.
  • A counterfactual is something that is contrary to fact.
  • In an experiment we observe what did happen when people received a treatment.
  • The counterfactual is knowledge of what would have happened to those same people if they simultaneously had not received treatment. An effect is the difference between what did happen and what would have happened.

Important

We could add a group of participants to a waiting list, do you have any example in mind?

Let’s finally define causal relationship

Shadish et al. (2002) :

  • This definition was first coined by John Stuart Mill (19th-century philosopher), a causal relationship exists if:
    1. The cause preceded the effect.
    2. The cause was related to the effect.
    3. We can find no plausible alternative explanation for the effect other than the cause.

Warning

Correlation does not prove causation!!! We will use this as a mantra in this class.

Let’s switch to important concepts

Don’t forget how variable is the concept of variable!

Independent variable

  • Independent variables: are those that influence, or affect outcomes in experimental studies. They are described as “independent” because they are variables that are manipulated in an experiment and thus independent of all other influences.
  • However, we will use this concept more vaguely, we won’t use it only when talking about experiments. It will be used also for correlational relationships, formally its name in survey designs is predictor.

Dependent variable

  • Dependent variables: are those that depend on the independent variables; they are the outcomes or results of the influence of the independent variables. It is also called outcome in survey designs or correlation designs.

Moderating variables

Moderating variables are predictor variables that affect the direction and/or the strength of the relationship between independent and the dependent variable.

Mediating variables

  • Mediating variables stand between the independent and dependent variables, and they transmit the effect of an independent variable on a dependent variable.

The art of forging hypotheses

What is a hypothesis?

  • The scientific method proposes that hypotheses are an important component to gather insights about nature.

  • A hypothesis is an statement regarding what we believe might be happening in nature, or beliefs about what has happened in nature.

  • A hypothesis can be for instance the belief that it rained at night when you wake up the next day and you see wet grass. You may create the statement:

“The last night it rain because I can see the grass is wet”

  • But wait, is there an alternative possible explanation? Yes there is:

“The last night I saw the grass wet, I believe the sprinklers were on last night”

  • These are observations about nature, and they can be hypotheses. Let’s make them more hypothesis-like:

“The grass is wet because it rained last night”

“The grass is wet because the sprinklers were turned on last night”

-These are statements that you can test and see if they are true or false. How? You may wake up earlier the next morning and check if the sprinklers were working.

  • But wait! One observation might not be enough, you don’t know at what time the sprinklers are working if they are really working. Then, you must wake up every morning at different times and record if the sprinklers were working or not. After several mornings, you may find that the sprinklers were a possible cause. Why possible? Because you still don’t know with this method, if the first time you saw the wet grass it rained. And here is where we need to precise different research methods.

What is a hypothesis? (2)

  • Talking about grass is not fun, and it is simplistic. It is better if we make more realistic hypotheses in psychology.

  • An example is happiness in married couples. Many researchers have found out that married people report more happiness compare to single people. See this article as example: Does marriage make people happy, or do happy people get married?

  • Let’s try to generate a hypothesis:

    • Happy people has a tendency to get married

  • Or we could state the following hypothesis:

    • Couples tend to be happier because they are married

  • We could also do the following statement:

    • There is a correlation between happiness and being married

  • If you read carefully you realized that all these statements are measurable, and they can be rejected. Also notice that the second hypothesis is a causal hypothesis, while the other two hypotheses are correlational hypotheses.

  • We need to create hypothesis that are falsifiable.

References

Creswell, J. W., & Creswell, J. D. (2017). Research design: Qualitative, quantitative, and mixed methods approaches. Sage publications.
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton, Mifflin; Company.