To understand the ferroelectric behavior, you have to study the Polarization vs Electric field character (PE loop). If the result carry the hysteresis loop then the material system consider as a ferroelectric.
You may follow this one ' Enhanced multiferroism in Gd-doped BiMn2O5 ceramics'- DOI: 10.1016/j.ceramint.2017.10.080 .
You can try the ferroelectric behavior of your samples using techniques such as the PE loops (which is the measurement of the charge displacement under a voltage dQ/dV). However, charge screening can provide wrong interpretation of the PE data. So you can also do a piezoresponse measurement using PFM and calculate the d33 coefficient (just in case you have a piezoelectric material).
Also, temporal electrical polarization (shock or disturbance) features give a higher impact in observing small(er) signals of an electrical polarization; the application of a small temporal electrical (and/or a mechanical in pzt-crystals) strain results in the generation of an electrical polarization (q-disturbance) across the material. A fast (ms to ns) removal of the potential (or a swithing) allows the ferroelectric (or a pzt-crystal) material to restore towards the initial disturbance to its original polarization (charge state and/or orientation). So, we might observe (in Z-impedance) hysteresis-loop(s), shown as a quasi-circular Z-loop(s), in a Zr-Zi (Nyquist) plot, or a near perfect Z-oscillation(s) in a Bode Z(f) plot[1].
1. Basics of a Quartz Crystal Microbalance https://www.gamry.com/application-notes/physechem/basics-of-a-quartz-crystal-microbalance/
Two properties coming inherently with ferroelectricity are piezoelectricity and pyroelectricity. If in doubt about ferroelectricity of some material, you might check for these properties: especially if pyroelectricity is missing, the theory of ferroelectricity has to be dismissed.
In addition to the answer of Dr. Dreher, I want to add one point
While measuring the PE loop, several spurious factors arise like deopolarizing field, FN tunneling, schottky effect which may affect your measurement. You can go through a paper by James Scott ' Ferroelectrics go Bananas' where it is shown that the PE loop you get may not be the instrinsic Polarization. We use an evolved protocol from ST circuit where we use remanant polarization. For a clearer view you can see my paper 'Multiferroicity around Verwey transition in Fe3O4 thin films.