Variables in Research PDF

## Variables in Research **Any object, quality or event that changes in some way.** | Variable | Description | Example | |---|---|---| | Independent variable (IV) | What the experimenter manipulates/controls in order to establish the relationship it has with the dependent variable. | Noise. | | De...

## Variables in Research **Any object, quality or event that changes in some way.** | Variable | Description | Example | |---|---|---| | Independent variable (IV) | What the experimenter manipulates/controls in order to establish the relationship it has with the dependent variable. | Noise. | | Dependent variable (DV) | The variable that is measured in an experiment. Not under the researcher's control. In Psychological experiments, this is usually the performance of the participants. Ideally, in a study the only thing that influences the DV is the IV. | Performance on the memory task. | | Extraneous variables | However, there are usually other factors that might affect the dependent variable and thus influence results. Researchers need to try and control these. | Time of day, mood, hunger, amount of sleep, previous experience of memory tests | | Confounding variables | If extraneous variables cannot be controlled and they interfere with the experiment, these would be considered variables that influence results. | Mood, amount of sleep, previous experience of memory tests | ## Advantages and Disadvantages of the Experimental Research Tool ### Advantages - Ideal when doing quantitative studies where variables are manipulated. - Can confirm or reject a hypothesis. - Can establish a relationship between variables. - Easier to comment on study's validity and reliability than other research tools used in qualitative studies. - Numerical data is easy to analyze. - The experimental method offers the best chance of objectivity. ### Disadvantages - When participants are aware that they are participating in an experiment, evaluation apprehension could occur i.e, participants feel concerned that their performance is being judged. - Need a large number of participants. - Not every subject can be investigated using an experimental tool (e.g, sensitive topics). - Individuality not valued as same experiment is repeated with every participant. ## How to Conduct an Accurate Experiment - Researcher manipulates an independent variable (IV). - Researcher measures the dependent variable (DV). - Researcher attempts to minimize the effects of any possible extraneous variables on the DV. - After the experimenter analyses the results he/she should then draw a conclusion about the relationship between the variables. - Experimenter confirms or rejects the research hypothesis. ## Example of a Research Using the Experimental Research Tool **Title:** Measuring the relationship between caffeine and stress. **Hypothesis:** Participants who are given a high dose of caffeine in coffee will increase their stress levels. **Null Hypothesis:** There will be no difference in the level of stress reported by participants who receive a high dose and those who receive no dose of caffeine. **Independent Variable:** Amount of caffeine in the coffee. **Dependent Variable:** Participants' stress levels. **Operationalisation:** Variables must be clearly defined so that they can be measured accurately, contributing to a more valid experiment. **Possible Extraneous Variables:** - Caffeine in drinks that participants have drunk before experiment. - The participants' expectations of the effects of coffee. - Feelings of stress from the experiment. - Inappropriate time to hold the experiment (too early or too late at night). - Participants' Moods. - Participants' thoughts about what happened before or after the experiment. - Heat inside experimental room. - Excessive noise inside experimental room. - Uncomfortable seating during the experiment. - Demand characteristics: participants' ideas about the purpose of the study that influences their behaviour such as wanting to please the researcher or deliberately doing the opposite (the screw-you effect). ## Of Which Possible Confounding Variables (variables research cannot control and will affect DV): - Caffeine in drinks that participants have drunk before experiment. - Participants' moods. - Participants' thoughts about what happened before or after the experiment. - Feelings of stress from the experiment situation itself. - The participants' expectations of the effects of coffee. - Demand characteristics (definition on pg 70). ## Operationalising Variables ### Operationalising Coffee In this example, the amount of caffeine needs to be clearly defined. For example, 5 cups of coffee will be given so that 250mg of caffeine will be ingested. ### Operationalising Stress Stress also needs to be clearly defined in order for it to be clearly measured. For example, it can be defined according to scores on a questionnaire which asks participants to rate their stress or conduct a physical stress test that monitors heartbeat and blood pressure. ## Different Types of Experiment Groups Since there could be other factors interfering with the participants' level of stress (and hence stress not increasing due to caffeine), the researcher needs to have 3 separate groups (treatment group, control group, and placebo control group): - **Participants in each group need to be as similar as possible.** For example, only including 30 year old Maltese females, so that you do not include extra confounding variables that might interfere with results (in this case age, ethnicity and sex are kept constant in each group of participants). - **Participants should be randomly assigned to these three groups:** - **Treatment/Experimental group:** Participants who are given some coffee. - **Control Group:** A matched group of participants who receive no treatment and act as a comparison to the experimental group to study any effects of the treatment. Participants who do not drink coffee. - **Placebo control group:** Participants who receive a placebo (fake) treatment in order to check whether there will be a change in behaviour based on the expectation or belief that a treatment will have certain effects. Participants who (unknowingly) drink decaffeinated coffee. ## Placebo Effect When a person responds to some stimulus as if it were having an expected effect when in fact it should have no effect at all. Imagine two groups of people, you give one group alcohol and another group a liquid that we tell them is alcohol. The placebo group who received a pretend alcoholic drink act as if they consumed alcohol. Hence, it is important to include a placebo group to evaluate the difference between the placebo effect and the actual effect of the intervention. Let’s imagine that all the three groups started with an average stress level of 65/100. After the study, average stress levels of the 3 groups changed as follows: - **Control group:** 70/100 (increase of 5%, due to stress of being in the study). - **Placebo Control group:** 75/100 (increase of 10% due to stress of being in the study 5% and their expectation that coffee will increase stress 5% ). - **Treatment group:** 85/100 (increase of 20% due to stress of being in the study 5%, their expectation that coffee will increase stress 5% and caffeine itself 10% ). At first glance, one would think that caffeine increased stress level by 20% (from 65 to 85). However, one needs to deduct the stress of being in the study (5% raise of control group) and the participants' expectations (another extra 5% noted in the placebo control group). Hence, in this case, one can conclude that caffeine increased stress levels by only 10%. **Conclusion:** So, if stress levels do increase in the participants after we give them coffee, we can reject the null hypothesis and accept the research hypothesis. ## Establishing a Relationship Between Coffee and Stress When conducting an experiment, the researcher's and participants' expectations can both interfere with the results. Therefore, it wise for the researcher to use the single- or double-blind procedures. - **Single-blind procedure:** Participants must not be aware about which group they are in, because otherwise this will affect the results. This reduces the effects of demand characteristics (definition on pg 70). - **Double-blind procedure:** Neither the experimenter nor the participants know to which group participants have been assigned to. The reason is that the researcher's expectations can also be a confounding variable and bias the results. ## True vs Quasi-experiments Experiments can be classified as either true or quasi experiments. ### True Experiments - This is thought to be the most accurate type of experimental research. - A true experiment supports or refutes a hypothesis using statistical analysis. - It is the only experimental design that can establish cause and effect relationships, when all variables are controlled. There are three criteria that must be met in a true experiment: 1. Researcher manipulates the independent variable. 2. Researcher includes a treatment as well as a control group. 3. Researcher randomly assigns participants to these 2 groups. #### Advantages - Researcher has a high level of control on variables. - Greater advantage in obtaining accurate results. #### Disadvantages - Individuality not valued as same experiment done with every participant. - Evaluation apprehension can occur i.e. participants are concerned that their performance is being judged. - **IF ONE OR MORE OF THESE CRITERIA IS NOT MET, IT WILL AUTOMATICALLY BE A QUASI EXPERIMENT** ### Quasi Experiments - Any experimental procedure that does not adopt random allocation of participants when setting experimental and control groups. - Might involve natural independent variables such as sex or age, so the IV cannot be manipulated. Often, in a quasi experiment participants' characteristics cannot be changed in the context of an experiment to see what relationship they might have with the dependent variable. This goes against the criteria of a true experiment. **Example:** Measuring the differences in music ability between boys and girls in a school setting. In this case the IV is the sex of the child, which obviously cannot be manipulated. #### Advantages - Participants' reactions are more likely to be genuine because these studies are not usually carried out in an artificial setting. - Less time consuming and cheaper since researchers are not involved in extensive pre-screening and randomization of groups. #### Disadvantages - Less/no control over the IV. - Low validity due to inability to randomly assign participants to groups. ### Laboratory Experiments - Take place in an environment designed to maximize control over extraneous variables (example: time of experiment, noise during the experiment) - This ensures the validity of the study. - **Example:** Pavlov’s classical conditioning experiment. #### Advantages - Offer consistent results and thus high reliability. - Easy to replicate (repeat) since lab experiments can be checked and confirmed by other researchers given that a standardised procedure is used. - Easy to establish a relationship between variables because extraneous variables are more easily controlled. #### Limitations - High level of control over variables makes the environment/task completed artificial or unrealistic. - Offers low ecological validity so results cannot be generalized to other situations other than the experimental situation. - Due to participants' expectations and demand characteristics (their awareness of what the researcher expects to find or how they should behave), participants may start to act differently from what they would normally act. **An example of a demand characteristic:** Hawthorne effect - The tendency of some people to work harder and perform better when they are participants in a study. This means that their recorded behaviour is artificially high, as they are trying harder than normal. This usually leads to invalid conclusions. ### Field Experiments - Conducted in the natural environment where behaviour normally occurs. - One still needs to compare the results of the experimental and control groups, to confirm a relationship between variables. - Although field experiments increase ecological validity, this brings with it disadvantages as the experimental procedure will be affected by extraneous variables. - Therefore, the researcher needs to be cautious when establishing a relationship between the IV and DV. - **Example:** If interested in discovering the level of students' stress during an exam, I conduct the experiment in a school. #### Advantages - Behaviour in a field experiment is more likely to reflect real life because of its natural setting. - Higher ecological validity than a lab experiment. - Less demand characteristics, especially when participants do not know they are being studied. #### Limitations - Greater opportunity for extraneous variables. These might influence and bias the results. - More difficult to replicate than a lab experiment. ### Natural Experiments - Studying the effects of a naturally occurring event. - The difference in the situation before and after the event takes place is the independent variable, while the change in behaviour is the dependent variable. So these factors make it automatically a quasi-experiment. - Researcher records the relationship between the natural IV and the DV. **Example:** Hodges et al (1989) compared the development of children who have been adopted, fostered or returned to their mothers with a control group of children who had spent all their lives in their biological families. Hence, researchers are not responsible for the manipulation of the IV (whether the child is adopted, fostered, returned to parents or always lived with biological parents). #### Advantages - Offer greater ecological validity as natural experiments are more realistic and avoid artificiality of laboratory experiments. - Allow researchers to study rare events that would be financially or ethically impossible to study otherwise. - Minimise demand characteristics as participants might not be aware they are being studied. #### Limitations - It is nearly impossible to replicate the experiment with the same conditions. - It is difficult to control confounding and extraneous variables increasing the chances of biased results. - May be more expensive and time consuming than lab experiments. ## Important Key Words About Experiments - **A prediction:** of what the researcher expects to discover in the study. - **Research hypothesis:** Usually that there will be a relationship between the independent and dependent variables. It needs to be highly specific and not vague so that it can be appropriately tested and either supported or refuted. - **Null hypothesis:** A hypothesis which states that there will be no relationship between variables. | Hypothesis | Description | Example | |---|---|---| | Research hypothesis | Usually that there will be a relationship between the independent and dependent variables. | There will be a relationship between loud noise (over 85 decibels) and performance of the memory task. | | Null hypothesis | A hypothesis which states that there will be no relationship between variables. | There will be no relationship between loud noise (over 85 decibels) and the performance of the memory task. |