Experimental Research

If we want to test a causal relationship between variables, that X causes Y (X->Y) then we can conduct experimental research.

The experiment

We have a voluntary group of people,we havent drawn a sample from a population. We randomly divide this group in two smaller ones and first test yb for both groups.

y (independent variable)= ability to memorize details from information

In the pre-test there were 33 variables. Instead of using them independently we calculate a scale score.

x= ability to mobilise or use prior knowledge

Each group was given a different text.

Then another test was given to us to memorize 20 words in a matrix. (post-test)

The x variable was not measured, it was used to create different circumstances for each group, different experimental treatment

----

The different circumstance was that for Group 1, the words in the matrix were already in the text that they were reading. That is because they already had a frame or structure that they could mobilize when they were trying to remember those words

The hypothesis here is that students of group 1 would perform better than those of group 2.

In order to subtract Ypre from Ypost (d = Ypost - Ypre) we have to standardize the values first because they are in different ranges. We do that via the Z-transformation.

Important Questions

insert sxima edw

Why do we say that this design is experimental?

because we didn't just observe and measure but we performed an action to alter the state of things, we created certain values

Why was randomization used?

The purpose/benefit was that the groups are in the same level at the start of the experiment.

If randomization is considered successfule, then the two groups should have equal mean in pre-test score.

if e.g people were categorized based on ages or gender it might have an effect

there are other external variables that might have an effect on the y variable, e.g

z1= how well you slept last night

z2=...

However if we randomize we try to create a barrier between the external variables and our system (that x->y, x causes y)

What else must be done in order to maximize the internal validity of the design?

Circumstances for both groups should be exactly the same, with the exception of the experimental treatment (x,~x)

for example we bring all participants in a laboratory setting, we bring them at the same time of the day, etc.

In educational research, true experiments are very rare.

Weaknesses of experimental research in education

ethical issues

There are instances that experimental research could harm the participants or hinder their progress, e.g by causes of violent behavior

ecological validity: participants have certain patterns of behavior,norms and established roles in their natural settings. If we take them out of those settings and put them in an unfamiliar, artificial environment, do we take results that can be generalized in their natural settings?
Hawthorne effect:

Two interpretations:

Participants behave differently when they know they are being observed

If the observation is long enough, it becomes a familiar part of the environment and they start behaving more naturally, but this takes time

If we change the light, e.g we dim the lights, and then the participants are asked to perform the same task , the productivity will increase. If we brighten the lights, it will also have the same effect. If e.g we introduce background music, the productivity will again increase.

The explanation is that humans are creatures of habit. We create certain routines and we are accustomed to our environment. If anything changes, we become more alert and our behavior changes.

This effect quickly wears off.

https://www.investopedia.com/terms/h/hawthorne-effect.asp

In our faculty, we do experimental research. In most cases, they are not true experiments.

⮜⮜

⮝⮝

⮞⮞

Page updated

Google Sites

Report abuse

Experimental Research

Related Reading

The experiment

Important Questions