Kinds of variable

1. A variable is anything in a study that can change from one person or one condition to the next.
2. Four kinds come up again and again: the independent variable, the dependent variable, co-variables and extraneous variables.
3. An experiment has an independent variable that is changed and a dependent variable that is measured.
4. A correlation has two co-variables and manipulates neither of them.
5. Every study has extraneous variables in it, and the researcher controls them as part of the procedure.
6. Operationalising a variable means writing down exactly how it is set up or measured, in words another researcher could follow.

The independent variable and how it is manipulated

1. The independent variable (IV) is the variable the researcher manipulates, which means deciding what each participant gets.
2. It needs at least two levels, called conditions, because a single condition leaves nothing to compare.
3. An independent variable can be manipulated in many ways, and these are the ones that come up most often.
   1. Give two different versions of the same task, such as learning a word list in silence and learning it with pop music playing.
   2. Change the amount of something, such as 30 seconds to learn the list against 90 seconds.
   3. Change the setting or the timing, such as testing pupils in a quiet classroom and then in a busy canteen.
   4. Add or remove a feature, such as showing one group a leading question and the other a neutral one.
4. One condition is often a control condition in which the independent variable is absent, which gives a baseline to compare against.
5. Where a difference already exists in the world, such as age group or whether someone plays an instrument, the researcher selects participants who already differ.
6. A manipulated independent variable supports a claim that the independent variable produced the change in the dependent variable.
7. Setting up neat conditions often makes the task artificial, which lowers ecological validity.

The dependent variable and how it is measured

1. The dependent variable (DV) is the outcome recorded from each participant.
2. A dependent variable has to come out as something that can be counted or measured, and these are the forms it most often takes.
   1. A score on a test, such as the number of words recalled out of 20.
   2. A time, such as the number of seconds taken to press a buzzer after a light comes on.
   3. A tally from an observation, such as how many times a child shares a toy in 10 minutes.
   4. A rating on a scale, such as how tired someone feels from 1 to 5.
   5. A count from a questionnaire or interview, such as the number of times a person says they check their phone at night.
3. Every participant is measured the same way, with the same materials and the same scoring rules.
4. Whether the measure captures the behaviour being studied is a question about validity.
5. Whether two researchers watching the same participant would write down the same number is a question about reliability, and a clear scoring rule is what makes them agree.
6. A numerical dependent variable can be compared directly and used to work out a mean or a range.
7. A number leaves out the reasons behind the behaviour, so a score of 8 out of 20 does not say why the other 12 words were lost.

Co-variables and how they are measured

1. Co-variables are the two variables measured in a correlation, and the researcher manipulates neither of them.
2. Each participant gives one number for each co-variable, so the data come in pairs.
3. A co-variable is measured in the same ways as a dependent variable, so a test score, a time, a tally or a rating scale are all possibilities.
4. Some co-variables are self-reported, such as hours of sleep last night, and some are recorded by the researcher, such as a maths test score out of 20.
5. The pairs are plotted on a scatter diagram, one axis for each co-variable and one point per participant.
6. Measuring co-variables allows the study of things that could never be manipulated ethically or practically, such as how much sleep a 15-year-old gets.
7. Because neither co-variable is manipulated, a correlation cannot support a claim about cause, since a third variable such as illness could be moving both numbers.
8. Self-reported co-variables carry a validity problem, since a pupil guessing at last night's sleep may be out by an hour either way.

Extraneous variables and how they are controlled

1. An extraneous variable is anything other than the independent variable that could change the dependent variable.
2. Participant variables are differences between the people, such as reading age, how much sleep they had or whether they already knew some of the words.
3. Situational variables are features of the setting, such as corridor noise, the time of day, the lighting or a different room.
4. Investigator effects are the researcher's own tone, wording or expression giving away what is expected.
5. Demand characteristics are cues that let a participant guess the aim, after which they may try to help or deliberately do the opposite.
6. An uncontrolled extraneous variable gives a second explanation for the results, so the change can no longer be put down to the independent variable, and validity drops.
7. Extraneous variables are controlled in a number of ways, and these are the ones used most often.
   1. Hold the variable constant, by running every participant in the same room at the same time of day with the same materials.
   2. Remove it, by using invented words so nobody can already know the list.
   3. Screen it out at the start, by only recruiting pupils who report a similar amount of sleep.
   4. Standardise the procedure, so that whatever the researcher says and does is identical every time.

Standardisation

1. Standardisation means every participant meets an identical procedure, so the only planned difference between them is the condition of the independent variable.
2. Standardised instructions are written out and read word for word, or printed and handed over, so nobody gets an extra hint.
3. The timings are fixed as well, such as 60 seconds to learn the list and two minutes to write it down.
4. Because the procedure is the same for everyone, a difference in the dependent variable is far more likely to come from the independent variable, which protects validity.
5. Writing that procedure down lets another researcher repeat it exactly and see whether the same results come out, which is how reliability is checked.
6. A tightly scripted session feels unlike ordinary life, so a result about learning invented words in a silent room may not transfer to revising at the kitchen table.
7. A rigid script can also make participants feel tested, which is the situation that produces demand characteristics.

Naming the variables in a study

1. Start with the thing being changed, which is the independent variable.
2. Write out both conditions in full, including the exact materials each one uses.
3. State the number recorded from every participant, which is the dependent variable.
4. Name one participant variable and one situational variable that could move that number.
5. Say how standardised instructions and a fixed setting would hold each of them still.

Turning this into marks

1. When an independent variable and a dependent variable are both asked for, give them as two genuinely separate things, one changed and one recorded.
2. Operationalise the dependent variable, such as words recalled out of 20, because "memory" does not secure the knowledge mark.
3. Name the extraneous variable and the control as two separate points, because corridor noise is the variable and testing everyone in the same quiet room is the control.
4. The mark most often lost is offering a control the study already runs, such as standardised instructions in a study that says they were read out.