As you adequately acknowledged the possibility of multiple factors that come to play, I would suggest you first eliminate some of these factors to justifiably accuse the elevation and rainfall at upstream point X to be responsible for flooding at downstream point Y.

Among the factors i would suggest taking care of include:

a) stream flow dynamics, including radius of curvature, discharge, velocity, bed flow characteristics, including mean depth;

b) Gradient between upstream point Y and downstream point Y or general topography of the region in question

c) Land cover change or general land use features at downstream point Y

d) Antecedent moisture at downstream point Y.

e) the amount, intensity and duration of rainfall at upstream point X coupled with surface characteristics / land cover/land use at the same point.

etc etc etc

Of importance would be to assess the extent to which these and many other factors would ordinarily influence the lead time of downstream inundation and or the spatial extent of inundated area.

Strangely, these features and many others may even frustrate flash floods but would cumulatively lead to riverine flooding.

Dr. Mwanang’ono gave a very good answer. I would like to hone in on a few Items, channel types and gradient, sediment loading. Gully stream types have moderate gradients, but so entrenched, no floodplain and limited sinuosity to dissipate flow velocity and delivery of rapid flow downstream. At another extreme, braided stream from excessive sediment have low gradient and high flooding frequency, but not especially deepest. Elevation change at a falls can be abrupt and extreme, but the energy to move water rapidly is somewhat dissipated at the plunge pool at the base of the falls, so the efficiency to move water downstream is somewhat limited as compared to a Rosgen A or B channel type that has high to moderate gradient and limited sinuosity and floodplain, with higher velocities, and able to deliver floodwaters downstream with efficiency. Gully networks also are expanded in extent, and they too deliver flow rapidly. Som of the most extreme examples of high flood peaks for unit area are on smaller catchments and watersheds, as rainfall extremes are more able to encompass the entire or most of the area for extended periods. Even hurricanes though large and extreme, are seldom able to encompass an entire river basin at one time. Perhaps conditions mentioned as atmospheric rivers can create the greatest extremes, or cyclonic typhoons or hurricanes or even thunderstorms that move slowly. Flash floods are also noted in some arid areas where the soils can develop hydrophobicity, vegetative cover minimal, and channel networks tend to be confined, entrenched.

Using a Geographic Information System (computing software and hardware made for answering questions about geography and hydrology), watershed maps and elevation maps compare watersheds known for flash flooding against watersheds that have had no flash flooding. You are going to bog down (pun intended) in things like the carrying capacity of the water system, the water retention of the soils, the tendency for retention dams to form and break and rain/snowfall history. Simply stated flatter elevations with silty, beaver habitat are more likely to release pulses of water than larger drops of elevation that produce rapid, fast clearing streams and rivers which have not deposited ages of soils.

If you are doing a science fair demo then two kid pools at different slopes with equal sized drain holes will show a correlation but reality is much more complex.

Get a GIS and compare a lot of different watersheds and run a stastical correlation.