Chi-square and Beyond for 2x2 Tables
18 February 2025, 9:30 am–5:00 pm
This course aims to demystify the range of statistics that can be used to summarise the associations between two binary variables. We move on from the standard chi-square test to investigate other options that may yield more useful information.
Event Information
Open to
- All
Availability
- Yes
Cost
- £125.00
Organiser
-
Centre for Applied Statistics Courses
Book Now (Hybrid)Note: This course will be delivered in hybrid mode, i.e. the in-person lecture will be transmitted live on zoom. The face-to-face part of the hybrid courses will go ahead conditional on at least 5 participants registering for this mode of delivery. Few classes may be delivered exclusively online, so please check under the Future dates tab for clarification. The online part of the class will be supported by its own dedicated teaching staff (along with the lead lecturer), therefore participants can expect the same level of group and individual support as the face-to-face class. You will be able to choose your preferred mode of attendance (face-to-face or online) during registration.
- Details
When data is collected from 2 binary variables on the same individuals or items, the data may be displayed as a 2 by 2 table. It is often of interest to determine whether there is some form of association between the two variables in the table, for example, comparing the % who test positive between two treatment groups, or between 2 subgroups such as males and females. The chi-square test is most commonly used by researchers to compare the percentage of individuals in two groups who share a trait.
However, there are many other statistics that may be applied to a 2x2 table to provide more useful information and this course aims to help guide the researcher through alternative analyses to decide on the most appropriate investigation of their data.
The following tests and concepts are covered in the course:
- Fishers exact testing
- Absolute Risk Reduction (ARR)
- Proportions tests
- Number needed to treat (NNT)
- Relative risk (RR)
- Attributable risk (AR)
- Odds ratio (OR)
- Kappa
- Sensitivity and specificity
- Positive and negative predictive values
- Likelihood ratios
For each statistic we consider the interpretation and appropriate usages.
- Fees
Course fees
External Delegates (Non-UCL) £125.00 UCL Staff and Students £62.50 * UCL Alumni £62.50 Staff and Doctoral Students from ICH/GOSH FREE † * Valid UCL email address and/or UCL alumni number required upon registration. Please note, this category does not include hospital staff unless you hold an official contract with the university.
† Limited free spaces available. If there are no free places remaining, Staff and Doctoral Students from ICH/GOSH can still register at the UCL rate.
Cancellations
Please note that no refunds will be given for non-attendance or cancellations made within 5 working days of the start of the course. For delegates attending courses on funded places, a £50 fee will be charged for late cancellation, non-attendance or partial-attendance.
- Future Dates
Dates Time Apply 18 February 2025 9.30am - 17.00pm Book Now ^We recommend logging in 10 minutes prior to the scheduled start time to access materials and ensure the course can start promptly at 9.30am.
- Feedback
“A really useful day. Just the right level for me and addressed some knowledge deficits. Lots to think about”
“Thank you very much, the course was indeed helpful!”
"I found this course extremely interesting and useful. It clarified many points of definitions and limitations of the different tests. I thought the practicals were essential and useful to consolidate the points learnt.”
"A great course, it brought together a lot of things, I didn't realise were connected.”
"The presenter was very good. It was good doing all the practicals to help understand the concepts.”
"The lecturer was great and very engaging.”