Login

Welcome, Guest. Please login or register.

March 19, 2024, 08:58:19 pm

Author Topic: Scientific Methodology Q&A  (Read 12024 times)  Share 

0 Members and 2 Guests are viewing this topic.

Bri MT

  • VIC MVP - 2018
  • Administrator
  • ATAR Notes Legend
  • *****
  • Posts: 4724
  • invest in wellbeing so it can invest in you
  • Respect: +3677
Scientific Methodology Q&A
« on: October 06, 2017, 11:17:48 am »
+17
Ask your questions about the scientific method here!

Recently there has been an influx of questions about scientific methodology, which given the recent introduction to the study designs of all subjects makes sense. We are going to be assessed on this, and it will be a significant portion of the exam. So we do need to prepare.

Below, I have written some general points which should help you get in the right direction.

Variable types
variable refers to the ability to change. Change in one variable can result in another one changing. In experiments, we measure and control these changes.

Dependent variable (DV): This is the variable which is being changed as a result of other variable in the experiment. (its value is dependent on the other variables)
-This variable belongs on the y-axis of a graph
-You want the changes to only be caused by the independent variable

Independent variable (IV): This is the variable which because it is changing, it impacts the dependent variable. (its value is independent of the values of the other variables)
-this value belongs on the x-axis of the graph
-You want this to be the only variable impacting the dependent variable/s

Extraneous variable (EV): This is a variable which might have an impact on the dependent variable
-you do not want these variables around
-get rid of them if you can
-keep them constant if you can (ie. make them  controlled variables  )
-you DO want to mention these in the discussion
-quantify their impact if you can
-if you cant control them, or they werent controlled in the experiment you looked at   definitely still talk about them
-the reason we don't like them is because it makes it hard for people to know if things happened because of the IV or because of EVs, which then makes it very difficult to draw conclusions about the relationship between the IV and DV

(this one I think is only looked at in psych) Confounding variable (CV): these are extraneous variables that DID impact the results
-if one of these comes up in the 10 marker TALK ABOUT IT, and explain how it impacts validity

Validity:
-Did you actually measure the impact of the IV on the DV or did you measure something else?
Reliability:
-Were the experiment repeated multiple times?
-Did you get the same or similar results each time?

Accuracy:
-Do your results reflect the actual truth?

Precision:
-How many decimal places or significant figures do you have?
-What's your uncertainty?
-For an analog instrument, have uncertainty as +/- half the distance between two adjacent markings   unless uncertainty is provided by manufacturer
-For an analog instrument, you can cite a value as being halfway between two adjacent markings

Uncertainty:
Systemic: The measured values are different to the actual values in the same way each time (eg. thermometer always measures 0.5 K higher than it should)
-usually due to instruments

Random: The measured values are different to the actual values in a different way each time (eg. one time measure the value as 0.5 mm more than reality, the next you measure it as 1.0mm less than reality)
-usually due to humans
-account for these by repeating multiple times and averaging,  that way the shifts to either side should balance each other out

Operations:
-if adding or subtracting values, add the uncertainties together to find the uncertainty of the new value
-if multiplying or dividing values, add the percentage/relative uncertainties together to find your new values
Chemistry students: I have been told this at a revis lecture by an examiner: if you are adding things together, don't base your answer on the number of sig figs, base it on the least number of decimal places (given that they all share the same unit).
 
On experiments
An experiment aims to determine information about the relationship between variables.
If it is a case study, it is NOT an experiment.

Graphs
-use a greylead and a ruler
- depending on the relationship, the line of best fit may have to pass through (0,0)
- values further away from each other provide less uncertainty in the gradient of the line of best fit, pay attention to these data points
- do not extrapolate
- select a scale which best displays the range of data you have, and figure out the scale you are using for the x and the y axis separately    (this is also important for methods)
-when reading a graph ALWAYS check the units
-Physics students:  when asked to find the experimental value of (thing) which happens to equal the gradient of the graph, make it clear that you are using the gradient of the line of best fit to find this

Hypothesis:
-This is a statement, not a question
-You need to have your IV, DV, and the expected relationship
-This should be based on accepted theory, wherever possible
-This needs to be testable (eg. It has to have the ability to be supported or not supported by the results)
-This should be discussed when talking about your results  (ie. how do the results support or not support the hypothesis)
-it is NEVER EVER "proven" right or wrong.
-Psych students: stop operationalising variables in a research hypothesis
This is a research hypothesis from VCAA psych exam 2013 "That adolescents experiencing increasing levels of sleep deprivation will show increasing irregularity in their sleep recovery patterns."



Best of luck! Please respond with any questions, and if you are unsure of anything mentioned here, say something :)

Dr. Nick

  • Trailblazer
  • *
  • Posts: 33
  • Hi everybody!
  • Respect: 0
Re: Scientific Methodology Q&A
« Reply #1 on: January 22, 2018, 07:06:50 pm »
0
This is brilliant!

TheAspiringDoc

  • Guest
Re: Scientific Methodology Q&A
« Reply #2 on: January 22, 2018, 08:03:56 pm »
+1
Great! Could you expand a bit on precision vs accuracy?

P.s.
Spoiler
when I first saw the title I thought it was Scientology Q&A
Edit: 800th
« Last Edit: January 22, 2018, 08:15:02 pm by TheAspiringDoc »

AngelWings

  • Victorian Moderator
  • ATAR Notes Superstar
  • *****
  • Posts: 2458
  • "Angel wings, please guide me..."
  • Respect: +1425
Re: Scientific Methodology Q&A
« Reply #3 on: January 22, 2018, 08:43:53 pm »
+3
Great! Could you expand a bit on precision vs accuracy?

P.s.
Spoiler
when I first saw the title I thought it was Scientology Q&A
Edit: 800th
Chem minor basically pounded this into me about a thousand times, so I guess it comes in handy right now.
Accuracy is how far away things are to the true result.
Precision is how many times you can get approximately the same result when you repeat the experiment.
Accuracy doesn't necessitate to precision and precision doesn't necessitate accuracy i.e. accuracy =/= precision.
The usual picture associated to this is a bullseye picture like the one on this link.
VCE: Psych | Eng Lang | LOTE | Methods | Further | Chem                 
Uni: Bachelor of Science (Hons) - genetics
Current: working (sporadically on AN)
VTAC Info Thread

randomnobody69420

  • Trailblazer
  • *
  • Posts: 49
  • Respect: 0
Re: Scientific Methodology Q&A
« Reply #4 on: February 18, 2018, 04:17:01 pm »
0
Hello, I'm having a bit of trouble understanding accuracy. What do you mean when you say the results don't reflect the actual truth? And what is accuracy dependent on (I'm guessing systematic errors is one of them but I could be wrong)? Thanks in advance!

Bri MT

  • VIC MVP - 2018
  • Administrator
  • ATAR Notes Legend
  • *****
  • Posts: 4724
  • invest in wellbeing so it can invest in you
  • Respect: +3677
Re: Scientific Methodology Q&A
« Reply #5 on: February 18, 2018, 05:01:44 pm »
+3
Hello, I'm having a bit of trouble understanding accuracy. What do you mean when you say the results don't reflect the actual truth? And what is accuracy dependent on (I'm guessing systematic errors is one of them but I could be wrong)? Thanks in advance!

Lets say that you're measuring the temperature of a lake. You might measure this 100 times (very reliable).  However, if your measurements indicate that the temperature of the lake is 25.76 C but it is actually 23 C your value lacks accuracy.

The most common analogy used is that of a bullseye. If your arrows are bunched together (similar values) this shows reliability/precision. If they are close to the bullseye (true value) this shows accuracy.

You're correct, systematic errors do negatively impact reliability.
Validity and accuracy also have a close relationship - an invalid experiment is highly unlikely to produce accurate values

Aksharan101

  • Fresh Poster
  • *
  • Posts: 1
  • Respect: 0
Re: Scientific Methodology Q&A
« Reply #6 on: March 30, 2022, 09:26:28 pm »
0
How do we increase accuracy and precision in an experiment?