I needed a quick refresh of ROC curves and this is exactly what I was looking for! Thank you for the clear explanation of how to perform it on SPSS Todd.
I have never used a ROC curve so it was cool to learn about it, and why we use it. At first I was confused on what the depression score was used for but when I looked back at the beginning, you said it was the "scale" which cleared things up for me on that.
Thank you so much! My course did not provide this very important information and expected me to be my own professor. I would not have understood any of my assignment if not for you.
This answered my question about finding the area, if it is above .9 the test has great specificity and sensitivity. Above .8 is considered good as well. Good to know!
Thank you so much for the clear explanation, it is really helpful! If i want to choose my own cut-off scores to look at eg. 1.5 instead of 1.55 given by SPSS, how may i do so?
Thank you Dr. Todd.. I'm doing this analysis on prognostic test. My question is, can it continue with PPV, NPV, LR+, LR- I mean, with this numerical data? Or should we categorized first? thank you
Dear Dr Todd Grande. Thank you very much for your RU-vid videos, especially on SPSS and mainly the ROC curve as they have helped me a lot. I am still struggling with how to generate a ROC table containing TP, TN, FP and FN given this(in this very video) particular form of ROC analysis where to predictor is set of continuous which performance and the cutoff for prediction of the "truth" are sought. Please how can I get SPSS to generate a ROC table containing TP, TN, FP and FN when doing ROC this way? Thank you.
Understanding how to balance the consequences of a cutoff score has to be determined by what is tolerated by the test. Is it the instrument being used that will determine if this is a good result to measure the test. High and low scores can be predictive of positive and negative outcomes.
Thank you for very informative video. I ahve question concerning signture with values. if you have different genes (they have values) and you want to combine to assess their diagnostic values, how does it work? thanks in advance
Thank you so much for your presentation, it is very useful. Is it possible to do a ROC cure using the data type you used for the sensitivity and specificity? i.e. diagnostic test and reliable test (the gold standard)
Dear Dr.Grande, Thanks for your video!! clearly explained as always! This video only include one independent variable and I would like to ask what should I do if my project have to classify cancer recurrence and non-recurrence using multiple genes? should I use the binary logistic regression or the linear discriminant analysis or else? I hope I can receive your PROFESSIONAL advice!!!
Thank you so much Dr. Grande! I need to calculate positive/negative predictive value, the % of correctly disease classification of an specific cut-off (previously calculated by your method in SPSS), would you mind to help me? Thanks!
Is it possible to change the cut-off values that are displayed in the 'Coordinates of the Curve'? The results are listed in .5 increments but my scale is in whole numbers.
Too complicated for me to follow without being motivated enough by deadlines later.. Interesting how there must be parallels between the way data is collected from people who have purchased hearing aids with the intent to use them to be analtysed and how data from someone recieving treatment for depression is similar in the way Windows would have had to be cusjtomized too when the data from the instrument being used to do the analysis. In the case of using an instrument to follow the progress of a mental health patient over time how can a person who owns a clinic with more than one professional working there relate all of the instrumental results in a clinic setting overall when it comes to best practice there if the 85% of relevance when using a diagnostic insturment is only valid when analyising the positive outcomes?